Abstract / Summary
Background: Sickle cell disease (SCD) is one of the most common severe monogenic disorders globally. SCD can cause severe pain, significant end-organ damage, pulmonary complications, and premature death. Red blood cell (RBC) transfusions are used to treat complications of SCD, e.g. acute chest syndrome (ACS), or they can be part of a regular long-term transfusion programme to prevent SCD complications.
Objectives: To summarise the evidence from Cochrane reviews of the benefits and harms of RBC transfusions versus no transfusion, or restrictive transfusion (to increase the total haemoglobin) versus liberal transfusion (to decrease the haemoglobin S level below a specified percentage), for treating or preventing complications experienced by people with SCD.
Methods: We searched the Cochrane Database of Systematic Reviews on 3 June 2026. We included Cochrane reviews of randomised or quasi-randomised controlled trials that addressed various SCD complications and had RBC transfusion as an intervention or comparator. We assessed the methodological quality of included reviews according to the AMSTAR (A MeaSurement Tool to Assess systematic Reviews) quality assessment. The primary outcomes of the overview were: mortality from any cause; SCD-related serious adverse events; and transfusion-related adverse events. We extracted GRADE assessments from included reviews.
Main results: We included 17 reviews, 11 had no included studies with an RBC transfusion intervention. Six reviews included trials with participants randomised to RBC transfusion; in one of these reviews, with no usable data, only 10 participants were randomised. This overview focuses on the five remaining reviews. Five reviews (containing nine trials with 1502 participants) reported data comparing short- or long-term RBC transfusions versus standard care; disease-modifying agents; a restrictive versus a liberal transfusion strategy; and long-term RBC transfusions versus transfusions to treat complications. All reviews were high-quality according to AMSTAR, however, the certainty of the evidence was variable across outcomes, with most being very-low certainty. Trials were downgraded according to GRADE methodology for risk of bias, indirectness (most trials were conducted in children with the HbSS phenotype), and imprecision (outcomes had wide confidence intervals). In all five reviews and all comparisons there may be little or no difference in the risk of death (low or very low-certainty evidence). There were either no deaths or death was a rare event. In all five reviews and all comparisons, evidence for several outcomes was of very-low certainty. These were: transfusion reactions, development of alloantibodies and serious infections. Short-term RBC transfusion versus standard care (one review: two trials, 434 participants) All reported outcomes were very-low certainty evidence. Long-term RBC transfusion versus standard care (two reviews: three trials, 405 participants) In children and adolescents at high risk of stroke, long-term RBC transfusions probably decrease the risk of stroke (RR 0.12, 95% CI 0.03 to 0.49; 2 trials, 326 participants; moderate-certainty evidence) and may decrease the risk of ACS (RR 0.24, 95% CI 0.12 to 0.48; 2 trials, 326 participants, low-certainty evidence) and painful crisis (RR 0.62, 95% CI 0.46 to 0.84; 2 trials, 326 participants; low-certainty evidence) compared to standard care. Long-term RBC transfusions may also decrease the risk of silent cerebral infarcts (SCI) in children with abnormal transcranial doppler (TCD) velocities (RR 0.11, 95% CI 0.02 to 0.86; 1 trial, 124 participants; low-certainty evidence), but there may be little or no difference in the risk of SCI in children with normal TCD velocities and previous SCI (RR 0.70, 95% CI 0.23 to 2.13; 1 trial, 196 participants; low-certainty evidence). In children and adolescents already receiving long-term RBC transfusions for preventing stroke, compared to standard care, continuing long-term RBC transfusions may reduce the risk of SCI (RR 0.29, 95% CI 0.09 to 0.97; 1 trial 77 participants; low-certainty evidence). In children with normal TCD velocities and SCI, RBC transfusions may increase the risk of iron overload (incidence rate ratio 14.42, 95% CI 5.41 to 875.17; 1 trial, 121 participants; low-certainty evidence). Long-term RBC transfusion versus RBC transfusion to treat complications (one review: one trial, 72 participants) In pregnant women, long-term RBC transfusions may decrease the risk of painful crisis compared to transfusion for complications (RR 0.28, 95% CI 0.12 to 0.67; 1 trial, 72 participants, low-certainty evidence). RBC transfusion versus disease-modifying agents (hydroxyurea) (two reviews: two trials; 254 participants) For primary prevention of stroke in children with abnormal TCD and no severe vasculopathy on magnetic resonance imaging or magnetic resonance angiography, who have received at least one year of RBC transfusions, there may be little or no difference in the risk of iron overload between RBC transfusion and disease-modifying agents (mean difference (MD) -1.80, 95% CI -5.16 to 1.56; 1 trial, 121 participants; low-certainty evidence); all SCD-related outcomes were very-low certainty evidence. For secondary prevention of stroke in children and adolescents, hydroxyurea with phlebotomy may increase the risk of painful crisis (RR 3.15, 95% CI 1.23 to 8.11; 1 trial, 133 participants; low-certainty evidence) and global SCD serious adverse events compared to RBC transfusion (RR 3.10, 95% CI 1.42 to 6.75; 1 trial, 133 participants; low-certainty evidence). There may be little or no difference in the risk of iron overload (low-certainty evidence). Restrictive versus liberal RBC transfusion strategy (one review: one trial; 230 participants) All reported outcomes were very-low certainty evidence.